Radiological features accompanying peroneus brevis split rupture revealed on magnetic resonance imaging – a cohort study

Study design

This study has a retrospective cohort design where already-existing MR images of the ankle were analyzed. The investigated parameters were compared using two groups: one consisting of patients with confirmed split tear of the PB tendon and one control group without a split tear. Additional inclusion and exclusion criteria were employed as described below. Our experimental group (Table 1) included 40 females and 40 males. The mean age was 50 ± 13 years. The right ankle was examined in 36 cases, the left one in 44 cases. The control group (Table 1) included 58 females and 57 males. The mean age was 40 ± 14 years. The right ankle was examined in 55 cases, the left one in 60 cases. Demographics of the study population included in the study are presented in Table 1. The average body mass index (BMI) was 24.1 kg/m2 in the experimental group and 23.7 in the control group (p > 0.05) (Table 2).

Inclusion criteria

MRI examinations of the adult patients (age > 18 years) performed between 2018 and 2021 (N = 239) at the Sahlgrenska University Hospital (SU) in Gothenburg, Sweden, were eligible.

A dedicated ankle coil was used for MRI acquisition at 3.0 Tesla. Only MRI examinations with the following sequences were included: proton density (PD)-weighted with and without fat suppression, T2-weighted and T1-weighted without fat suppression.

All ankle MRI examinations included PD-weighted turbo spin echo (TSE): echo time (TE) 45 ms and repetition time (TR) 2800–5000 ms; T2-weighted (TSE): TE 60 ms and TR 3000–5000 ms; and T1-weighted: TE 11.5 ms, TR 700–750 ms. Voxel size was 0.45 × 0.53 × 3.0 mm, slice thickness 3 mm and field of view (FOV) 14 cm.

The foot and ankle position during the examination

The patient was placed in a supine position. Ankle joint localization was maintained using a dedicated coil suited to the shape of the ankle and foot. To further lock the ankle and foot position, elastic wedge-shaped cushions were used.

Exclusion criteria

Exclusion criteria were recent fracture (n = 2 from the control group), neoplasm (n = 2 from the control group), sequences without angled axial projections (n = 6 from the control group and 8 from the experimental group), artifacts obstructing evaluation (e.g., metal artifacts, n = 7 from the experimental group and n = 4 from the control group) and conditions which severely altered the appearance of the ankle (n = 13 from the control group and 2 from the experimental group). In total, we excluded 44 cases (17 from the experimental group and 27 from the control group), Fig. 1. These criteria were chosen to include images that allowed for proper evaluation of the peroneus tendons and minimized interfering noise from other conditions and artifacts.

Fig. 1figure 1

Flow chart demonstrating how magnetic resonance examinations were collected. MRI – magnetic resonance imaging, RIS/PACS – Radiological Information System/Picture Archiving and Communication System

Assignment to experimental and control groups

Peroneal split tear was defined as a radiologically and clinically proved identifiable longitudinal tear of the PB tendon. All patients that met the inclusion and exclusion criteria were included in the split tear group.

The control group was assembled in a similar way. Matching exclusion criteria were applied but, in contrast, the image sequences were evaluated to not include any PB split tears.

Other variables included and evaluated

The soft tissue abnormalities included in the study were abnormality in the SPR, ATFL or CFL. Abnormalities of SPR were classified as total rupture, relaxed or thickened. Abnormalities of the ATFL and CFL were assessed as one of three grades [20]: grade 1 is interstitial ligament injury which manifests on PD-weighted or fat-suppressed PD-weighted MR images as mild intraligamentous signal hyperintensity, ill-definition of the ligament and pericapsular edema; grade 2 was assessed as a focal ligament fiber; grade 3 was complete ligament fiber discontinuity [20].

The anatomical variations included in the study were the low-lying PB, os peroneum, prominent peroneal tubercle and malleolar groove shape. The presence of os peroneum was defined as a separate bone situated within the PL tendon, typically in the area inferior to the os cuboideum. If the length of the peroneal tubercle, measured from the lateral margin of the calcaneus, exceeded the width of each peroneal tendon it was defined as prominent. Three malleolar groove shapes were included: concave, flat and convex. The PB was considered low-lying if the muscle belly was seen at least at the level of the lateral malleolus at the level where the posterior talofibular ligament (PTFL) attaches. The shape was assessed at the level of where the PTFL attaches to the fibula. Using measurement tools in the Radiological Information System/Picture Archiving and Communication System (RIS/PACS), a straight line could be drawn through the groove. If the groove curved towards the center of the fibula, away from the measurement line, the groove was considered concave. If the groove neither curved nor protruded from the measurement line, it was considered flat. If the groove protruded beyond the measurement line, it was considered convex.

To more accurately evaluate localization of bone marrow edema, the tibia, lateral malleolus, talus and calcaneus were divided into sectors (see Fig. 2). The distal tibia was divided into three parts: the medial malleolus (Tibia C), the lateral half (Tibia A) and the medial half (Tibia B). The fibula was divided into an anterior half (Fibula A) and a posterior half (Fibula B). The talus was divided both in terms of superior (Talus C)/inferior (Talus D) and medial (Talus B)/lateral (Talus A). The calcaneus was divided into a lateral half (Calcaneus A) and medial half (Calcaneus B).

Fig. 2figure 2

Illustrating the defined distribution areas of bone marrow edema. Bone section coded as described in Sect. 2.6. A, D, G – proton density-weighted images; B, C, E, H – T2-weighted images with fat suppression; F – T2-weighted image

No gap between the talus and the adjacent fat pad in the anterior joint recess was defined as no effusion (coded as Effusion 1). Fluid creating a gap between the talus and the adjacent fat pad in the anterior joint recess, as well as fluid expansion of the posterior recess, was defined as moderate effusion (coded as Effusion 2). If fluid also created a gap between the anterior tibia and adjacent fat pad, it was defined as severe effusion (coded as Effusion 3). Presence of the synovitis in the talocrural joint was noted.

We measured patients’ height and weight and BMI was calculated. History of trauma was considered if the patient had had an ankle injury in the last 6 months.

Observers

All MRI examinations were evaluated using the radiology information system/ picture archiving and communication system (RIS/PACS) of the hospital (AGFA©). The data collection procedure was done separately by observers who were radiologists with 12 (PS, observer 1) and 3 years’ experience (AB, observer 2 and SB, observer 3) and a medical student (MH, observer 4) who did master’s degrees in medicine at the Department of Musculoskeletal Radiology at SU in Gothenburg, Sweden. PS and AB evaluated the soft tissue lesions. PS, SB and MH graded the bone lesions. Before scoring, the most experienced radiologist (PS) held a training session with all observers to ensure conformity in evaluation. The final decision was made by consensus.

Statistical analysis

The proportion and percentage of feature occurrence, compared to the total number of patients in each group, was calculated (i.e., the probability of a feature occurring in each group). The analyses are based on comparisons of the probability of feature occurrence in both groups. This supposition was made due to the relatively large size of the groups. In this case, the probability was determined by analyzing feature frequency in the study groups. Binominal proportion confidence intervals (CI, 95%) were calculated using asymptotic normal approximation and Wilsons’s test, to evaluate variance. The Bonferroni-Holm method (B-H) [21] was used to adjust for multiple comparisons. The data was nominal, in sets of two unpaired groups. Hence, to determine the significance and magnitude of percentual differences, p-values and CI (95%) were calculated using two-sided Fisher’s exact test (FT). Statistical significance was always defined as p < 0.05.

Continuous variables (BMI) were found to be normally distributed. Parametric tests were used for variables with normal distribution. Differences in mean values were analyzed with Student’s t-test. A p value ≤ 0.05 was considered as statistically significant.

To investigate bivariate correlations, Pearson correlation for binominal variables was used to analyze the direction, strength (Pearson’s correlation coefficient, r) and significance of correlations.

To evaluate the inter-rater reliability of the method, the results from the two parties were compared using Cohen’s kappa coefficient for each parameter using SPSS. Cohen’s intervals were used for interpretation: no agreement (≤ 0), none to slight (0.01–0.2), fair (0.21–0.4), moderate (0.41–0.6), substantial (0.61–0.8) and almost perfect (0.81–1.0) [22]. In addition to Cohen’s kappa, cross-tabulation of percentual agreement was calculated for each parameter. Cohen’s kappa measured the degree of inter-rater reliability while the percentual agreement examined what differed in the evaluation.

Analyses and data visualization were performed using Microsoft Excel©, Statistical Package for the Social Sciences program (SPSS)© and Python 3.7.10 (with plotly 4.4.1, scipy 1.4.1 and statsmodels 0.10.2 libraries) in Jupyter notebook.

Ethics.

The ethical considerations of this study were primarily the management of confidential patient material and the digital storage of said data. Patients were assigned pseudo-encrypted codes through which deidentification was achieved. The Swedish Ethical Review Authority approved the study and waived the need for informed consent (number 2020-06-177 and 2021–05447).

The study was conducted in compliance with the Declaration of Helsinki. The anonymization of patient data ensured data protection following the European General Data Protection Regulation. The data were recorded in a password-protected secure database.

留言 (0)

沒有登入
gif